ExperimentSubset

ExperimentSubset implements data container classes for efficient storage and streamlined retrieval of row and/or column subsetted assays from Bioconductor Experiment objects while maintaining data provenance relationships between subsetted and parent assays.

Key Features:

• Efficient subset storage: Stores multiple subsets of the same data without creating separate objects for each new subset, reducing memory footprint.
• Row and column subsetting: Supports subsetting assays by rows (e.g., selecting most variable features) and/or columns (e.g., excluding poor-quality samples).
• Data provenance tracking: Maintains inherent relationships between subsetted assays and their parent assays to preserve data lineage.
• Streamlined retrieval: Enables efficient access to stored subsets at different stages of analysis workflows.

Scientific Applications:

• Single-cell RNA sequencing analysis: Stores and retrieves subsets at different stages of single-cell RNA-seq analysis while maintaining lower memory usage.
• Quality control filtering: Excludes poor-quality samples from assay matrices by column subsetting prior to downstream analytical methods.
• Feature selection: Selects the most variable features by row subsetting to focus analyses on biologically informative genes.

Methodology:

The package implements container classes that avoid copying original assay data when creating subsets, instead maintaining references to parent assays along with subsetting indices for rows and columns.